The circulation in the upper 2000 meters or so, beneath the Ekman layer and driven by the winds, is driven by convergence and divergence in the Ekman layer which provide Ekman pumping or suction to the interior ocean. This drives equatorward or poleward flow, respectively, since the mechanism whereby the slow, large-scale open ocean responds to changes in potential vorticity (essentially angular momentum) is by a change in latitude of the water column, such that its vorticity due to the earth's rotation is decreased or increased. This vorticity balance is referred to as Sverdrup balance [ Sverdrup, 1947]. With the discovery that the return flow must be on the western boundary in a slightly viscous ocean [ Stommel, 1948; Munk, 1950] the basic reason for the large, asymmetric subtropical and subpolar gyres was understood. However, it has been noted for decades that the western boundary current transports are larger than the Sverdrup transport integrated zonally across the ocean basin, primarily because of the existence of vigorous recirculations of the western boundary currents.
The Sverdrup balance in mid-basin has been checked recently using reference level calculations from hydrographic sections [ Schmitz et al., 1992, for the North Atlantic; Hautala et al., 1994, for the North Pacific]. In the North Pacific, the balance across the entire basin is surprisingly close, although there are two large-scale regional disagreements; that in the western Pacific may be accounted for by a barotropic component (that is, uniform with depth) of the Kuroshio which is not seen in the vertical shear. That in the eastern North Pacific appears to be either an unexplained additional gyre or possibly a time-dependent response to El Niño. Mid-latitude open ocean adjustment to changes in large scale forcing occurs through generation of Rossby waves, which have wavelengths of tens to thousands of kilometers and whose basic physics is a vorticity (angular momentum) balance between vorticity changes due to changes in latitude and due to either vertical stretching of the water column or generation of local vorticity. Jacobs et al. [1994] showed that the 1982-83 El Niño might have caused a Rossby wave to propagate westward from the eastern boundary of the North Pacific, which would have created a high surface height in the eastern Pacific in 1985, the year of the hydrographic observations.
The time dependence of the Sverdrup transport and actual transport has been studied recently for the Pacific. The Sverdrup transport is calculated directly from the wind stress. For the currents, usually only the relative baroclinic transports are given, that is, transport relative to a chosen level of no horizontal motion, since it is difficult to compute the absolute transport given measurement of the density field alone, which is usually the only measurement available. The Kuroshio transport has a large seasonal signal [ Sekine and Kutsuwada, 1994], with maximum baroclinic transport in the summer; the maximum Sverdrup transport calculated from the winds occurs in winter. Time dependence of the South Pacific subtropical gyre circulation is being explored by Morris and Roemmich [1994] using six years of repeated XBT sections between New Zealand and Fiji. They find a seasonal cycle, with a maximum in baroclinic transport in October, while the maximum in Sverdrup transport is in August each year. The baroclinic transport increase is accompanied by a shift of the shape of the circulation. While in both the North and South Pacific the baroclinic current changes lag the Sverdrup transport changes, as would be expected for westward propagation of the response, in neither case is the measured time lag understood: it lies between the barotropic response time (about a month) and the baroclinic response time (years). Part of the discrepancy may be due to unmeasured barotropic changes, as was suggested by Hautala et al. [1994] for the difference between the annual mean Sverdrup and Kuroshio transport in the western Pacific.